Calculate the TPS of the GeForce RTX 4090 Max-Q on local AI models

NVIDIA 16 GB GDDR6 576 GB/s January 2023

Every model in our catalogue assessed against this card at the context length and minimum quality you choose. Speed is an estimate for a single request, calculated from this card's memory bandwidth and the size of each model once compressed.

Calculated for this card

432 models it can run

679 models in our catalogue altogether

Largest model it holds

Nemotron 3-Nano-30B-A3B

31.6B · Q3_K_M · 116 tok/s

Fastest model

Gemma 3 QAT 1B

244 tok/s · 1B

Which AI models can run on a GeForce RTX 4090 Max-Q?

Set the inputs, read the answer

More context means more memory for the conversation cache. Speed is for a fresh conversation and does not change with this setting.

Hides models that would only fit by being compressed below this point.

432 models match

Calculating
Quantisation Fit
244 tok/s

207–293

Gemma 3 1B 1B Mar 2025 1.8 GB 33k tokens Q8_0 Comfortable
244 tok/s

207–293

Gemma 3 QAT 1B 1B Apr 2025 1.8 GB 33k tokens Q8_0 Comfortable
244 tok/s

146–390 · low confidence

HGRN 1B (WT 103) 1B Nov 2023 1.8 GB 131k tokens ? Q8_0 Comfortable
244 tok/s

146–390 · low confidence

LLama 3..2 Typhoon 2 1B 1B Dec 2024 1.8 GB 131k tokens ? Q8_0 Comfortable
244 tok/s

146–390 · low confidence

OLMo-1B 1B Feb 2024 1.8 GB 131k tokens ? Q8_0 Comfortable
244 tok/s

146–390 · low confidence

Pythia-1b 1B Apr 2023 1.8 GB 131k tokens ? Q8_0 Comfortable
226 tok/s

136–361 · low confidence

OpenELM-1.1B 1.1B May 2024 1.9 GB 131k tokens ? Q8_0 Comfortable
222 tok/s

133–355 · low confidence

DeciCoder-1B 1.1B Aug 2023 1.9 GB 131k tokens ? Q8_0 Comfortable
222 tok/s

133–355 · low confidence

SantaCoder 1.1B Jan 2023 1.9 GB 131k tokens ? Q8_0 Comfortable
222 tok/s

133–355 · low confidence

TinyLlama-1.1B (1T token checkpoint) 1.1B Oct 2023 1.9 GB 131k tokens ? Q8_0 Comfortable
222 tok/s

133–355 · low confidence

TinyLlama-1.1B (3T token checkpoint) 1.1B Oct 2023 1.9 GB 131k tokens ? Q8_0 Comfortable
203 tok/s

122–325 · low confidence

EXAONE 4.0 (1.2B) 1.2B Jul 2025 2.0 GB 131k tokens ? Q8_0 Comfortable
203 tok/s

122–325 · low confidence

MinerU2.5 1.2B Sep 2025 2.0 GB 131k tokens ? Q8_0 Comfortable
203 tok/s

122–325 · low confidence

Pleias 1.0 1.2B 1.2B Dec 2024 2.0 GB 131k tokens ? Q8_0 Comfortable
203 tok/s

122–325 · low confidence

Pleias-RAG-1B 1.2B Apr 2025 2.0 GB 131k tokens ? Q8_0 Comfortable
198 tok/s

169–238

Llama 3.2 1B 1.2B Sep 2024 2.2 GB 131k tokens Q8_0 Comfortable
196 tok/s

117–313 · low confidence

MiniCPM-1.2B 1.2B Jun 2024 2.0 GB 131k tokens ? Q8_0 Comfortable
188 tok/s

113–300 · low confidence

DeepSeek Coder 1.3B 1.3B Jan 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
188 tok/s

113–300 · low confidence

DeepSeek-VL-1.3B 1.3B Mar 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
188 tok/s

113–300 · low confidence

DigiRL 1.3B Jun 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
188 tok/s

113–300 · low confidence

GLA Transformer 1.3B 1.3B Aug 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
188 tok/s

113–300 · low confidence

Janus 1.3B 1.3B Oct 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
188 tok/s

113–300 · low confidence

Kosmos-2.5 1.3B Aug 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
188 tok/s

113–300 · low confidence

Otter 1.3B May 2023 2.1 GB 131k tokens ? Q8_0 Comfortable
188 tok/s

113–300 · low confidence

Phi-1 1.3B Oct 2023 2.1 GB 131k tokens ? Q8_0 Comfortable

Speeds are estimates for a single request — one conversation at a time — calculated from memory bandwidth, model size and quantisation. Real throughput varies with the inference runtime and its version. Figures published by hardware vendors measure many simultaneous requests and are much higher.

On record

GeForce RTX 4090 Max-Q full specification

Everything on record for this board, ordered by how much it bears on running a language model rather than by how a spec sheet would list it. Memory comes first because it decides the outcome; the rest is context.

Memory

The two specifications that decide what this card can run and how quickly. Capacity sets which models fit; bandwidth sets how many tokens per second they produce once they do.

Memory size
16 GB
Memory bandwidth
576 GB/s
Memory type
GDDR6
Memory bus width
256 bit
Memory clock
2.25 GHz

The chip

Which processor is on the board and how it was manufactured. A smaller process size generally means more performance for the same power.

Graphics processor
AD103
Architecture
Ada Lovelace
Generation
GeForce 40 Mobile
Foundry
TSMC
Process size
5 nm
Transistors
45.9 billion
Transistor density
121,100 K/mm²
Die size
379 mm²
Released
3 January 2023

Clock speeds

How fast the processor runs. Worth far less here than on a gaming benchmark: generating text is limited by memory bandwidth, so a higher clock barely moves the result.

Base clock
930 MHz
Boost clock
1.46 GHz

Processing units

What the chip contains. These drive graphics performance and matter mainly for processing a long prompt rather than for producing the answer.

Shading units
9,728
Texture mapping units
304
Render output units
112
Streaming multiprocessors
76
Tensor cores
304
Ray tracing cores
76
L1 cache
128 KB
L2 cache
64 MB

Theoretical performance

Peak arithmetic rates published for the board. These are ceilings that no real workload reaches, and generating text reaches a small fraction of them because it is limited by memory rather than arithmetic.

Half precision (FP16)
28.3 TFLOPS
Single precision (FP32)
28.3 TFLOPS
Double precision (FP64)
442.3 GFLOPS
Pixel rate
163 GPixel/s
Texture rate
442 GTexel/s

The board

What it takes to physically install and power the card — the practical constraints that decide whether it fits the machine you already own.

Power draw (TDP)
80 W
Power connectors
None
Bus interface
PCIe 4.0 x16
Slot width
IGP

Software support

Which graphics and compute interfaces the card supports. CUDA compute capability is the one that bears on inference: below 7.0 there are no tensor cores, and modern inference software falls back to slower code paths.

CUDA compute capability
8.9
DirectX
12.2
OpenGL
4.6
Vulkan
1.4
OpenCL
3.0
Shader model
6.8

Listings

Where to buy a GeForce RTX 4090 Max-Q

No vendor is currently listing this card. Listings come from vendors who publish them here directly — browse the vendor directory to see who is selling what.

What the numbers mean

Memory: the specification that decides everything

Memory

16 GB

Bandwidth

576 GB/s

Largest model

Nemotron 3-Nano-30B-A3B

16 GB of GDDR6 puts the GeForce RTX 4090 Max-Q comfortably into small and mid-sized models, with roughly 14.4 GB usable once the driver overhead is taken out. The largest models are out of reach without splitting them.

The memory bus moves 576 GB/s across a 256-bit bus. That is the number that governs generation speed — arithmetic per byte read is small enough that the bus, not the cores, is what everything waits on.

Bandwidth is clock times bus width, and this card clocks its memory at 2.25 GHz. Both halves matter, and neither is visible in a gaming benchmark.

Put together, the largest model that fits is Nemotron 3-Nano-30B-A3B at 31.6B, running Q3_K_M and producing around 116 tokens per second.

The chip and how it was built

The GeForce RTX 4090 Max-Q is built on the AD103 graphics processor, using NVIDIA's Ada Lovelace architecture, as part of the GeForce 40 Mobile generation.

The chip is manufactured by TSMC, on a 5 nm process, with a die measuring 379 mm², holding 45.9 billion transistors. A smaller process generally means more performance for the same power, though for language models it matters far less than the memory subsystem.

It was released in January 2023, roughly 3 years ago. Inference software support tends to follow hardware by a year or two, so a card of this age generally has mature, well-optimised code paths available to it.

Compute throughput, and why it matters less than it looks

FP16

28.3 TFLOPS

FP64

442.3 GFLOPS

Tensor cores

304

On paper the GeForce RTX 4090 Max-Q reaches 28.3 TFLOPS at half precision and 28.3 TFLOPS at single precision. These are peak figures no real workload sustains, and generating text reaches only a small fraction of them — decoding is limited by memory rather than arithmetic, which is why a card can look enormously powerful here and still produce tokens at an ordinary rate.

Double-precision throughput is 442.3 GFLOPS. It has no bearing on running a language model — no inference runtime uses it — but it separates datacentre parts from consumer ones, since the latter deliberately restrict it.

The card carries 304 tensor cores across 76 streaming multiprocessors. These accelerate the matrix arithmetic at the heart of a transformer, and they are what make prompt processing — reading a long document before answering — dramatically faster than it would otherwise be.

Clocks run from 930 MHz at base to 1.46 GHz boosted. Worth far less here than on a gaming benchmark: raising the clock speeds up the arithmetic, and the arithmetic is not what generation is waiting on.

Cache and processing units

The GeForce RTX 4090 Max-Q has 128 KB of L1 cache, backed by 64 MB of L2. Cache absorbs a share of the memory traffic that would otherwise hit the main bus, which is the one place on this page where a number other than bandwidth quietly affects generation speed — a large L2 lets more of the working set stay close to the cores.

There are 9,728 shading units, 304 texture mapping units, and 112 render output units. These drive graphics workloads and contribute to prompt processing, but they sit idle for much of the time a model spends generating a reply.

Power, size and installation

Power draw

80 W

The GeForce RTX 4090 Max-Q is rated at 80 W. Running a language model keeps a card busy in bursts rather than continuously — it draws hard while generating and idles between requests — so sustained draw over a working day is usually well below the rated figure.

The board occupies a igp. Worth checking against the case and power supply already in the machine, since the largest cards need considerably more of both than a typical desktop provides.

It connects over PCIe 4.0 x16. The interface governs how quickly a model is loaded from disk into the card, not how fast it runs once there, so a narrower link costs a few seconds at startup and nothing thereafter.

The extremes

The largest AI models that run on a GeForce RTX 4090 Max-Q

The biggest open-weight models that fit on this card, newest first. Each is shown at the best compression the card can hold.

  1. 01 North Mini Code 30B · Q3_K_M · Jun 2026 122 tok/s
  2. 02 Qwen 3.6-27B 27B · Q3_K_M · Apr 2026 24.4 tok/s
  3. 03 Qwen3.5-27B 27B · Q3_K_M · Feb 2026 24.4 tok/s
  4. 04 Nemotron 3-Nano-30B-A3B 31.6B · Q3_K_M · Dec 2025 116 tok/s
  5. 05 Nomos 1 30B · Q3_K_M · Dec 2025 122 tok/s
  6. 06 C2S-Scale 27B · Q3_K_M · Oct 2025 24.4 tok/s
  7. 07 Gemma-SEA-LION-v4-27B-IT 27B · Q3_K_M · Aug 2025 24.4 tok/s
  8. 08 ERNIE-4.5-VL-28B-A3B 28B · Q3_K_M · Jun 2025 131 tok/s
  9. 09 Qwen3-30B-A3B 30B · Q3_K_M · Apr 2025 122 tok/s
  10. 10 Gemma 3 QAT 27B 27B · Q3_K_M · Apr 2025 24.4 tok/s

The fastest AI models on a GeForce RTX 4090 Max-Q

Where this card produces tokens quickest. Smaller models dominate here, because generating each token means reading the whole model out of memory once.

  1. 01 Gemma 3 QAT 1B 1B · Q8_0 · 1.8 GB 244 tok/s
  2. 02 Gemma 3 1B 1B · Q8_0 · 1.8 GB 244 tok/s
  3. 03 LLama 3..2 Typhoon 2 1B 1B · Q8_0 · 1.8 GB 244 tok/s
  4. 04 OLMo-1B 1B · Q8_0 · 1.8 GB 244 tok/s
  5. 05 HGRN 1B (WT 103) 1B · Q8_0 · 1.8 GB 244 tok/s
  6. 06 Pythia-1b 1B · Q8_0 · 1.8 GB 244 tok/s
  7. 07 OpenELM-1.1B 1.1B · Q8_0 · 1.9 GB 226 tok/s
  8. 08 TinyLlama-1.1B (1T token checkpoint) 1.1B · Q8_0 · 1.9 GB 222 tok/s
  9. 09 TinyLlama-1.1B (3T token checkpoint) 1.1B · Q8_0 · 1.9 GB 222 tok/s
  10. 10 DeciCoder-1B 1.1B · Q8_0 · 1.9 GB 222 tok/s

Step by step

How to work out the tokens per second of a GeForce RTX 4090 Max-Q

You do not have to calculate anything by hand — the gputps.com calculator on this page has already worked it out for every model this card can hold. Reading off the answer takes six steps.

  1. 01

    Find the model in the table

    Every one of the 432 models this GeForce RTX 4090 Max-Q runs is in the table above. Search narrows it by name or by size.

  2. 02

    Match the context to your work

    Longer conversations cost memory on top of the weights. With 16 GB to work in, that is frequently the difference between a model fitting and not.

  3. 03

    Set a minimum quality if you need one

    Each model is shown at the best compression this card can hold. A minimum quality hides the ones that only fit by being squeezed further than you would accept.

  4. 04

    Read the speed and the range

    Speeds come with error bars for a reason. The best case here is 244 tok/s on Gemma 3 QAT 1B, and which inference software you use moves that by thirty to fifty per cent.

  5. 05

    Check the headroom before you decide

    A tight fit runs but leaves no room to raise the context later; comfortable has headroom. The memory column shows what each model needs against the 16 GB available.

  6. 06

    Open the model to compare cards

    Following a model through to its own page lists all the hardware that can run it, so you can see where the GeForce RTX 4090 Max-Q sits against the alternatives.

Answers

GeForce RTX 4090 Max-Q — common questions

01

How many tensor cores does a GeForce RTX 4090 Max-Q have?

The GeForce RTX 4090 Max-Q has 304 tensor cores across 76 streaming multiprocessors. They accelerate the matrix arithmetic a transformer is built from, which mainly speeds up processing a long prompt rather than producing the reply.

02

Does the GeForce RTX 4090 Max-Q support CUDA?

Yes. The GeForce RTX 4090 Max-Q reports CUDA compute capability 8.9. Capability 7.0 and above has tensor cores, which modern inference software uses; below that it falls back to slower code paths for quantised models.

03

What bus interface does the GeForce RTX 4090 Max-Q use?

It uses PCIe 4.0 x16. This governs how fast a model is loaded onto the card rather than how fast it runs once loaded, so it costs a few seconds at startup and nothing during generation.

04

Is the GeForce RTX 4090 Max-Q good for running local AI models?

Its memory covers small and mid-sized models, though the largest are out of reach and its bandwidth gives usable, if unspectacular, generation speeds. In total it runs 432 of the models we track. Whether that is enough depends entirely on which model you want — the table above answers that directly.

05

Can a GeForce RTX 4090 Max-Q run a model that does not fit in its memory?

Only partly. Layers beyond the 16 GB sit in system memory and run at a fraction of the speed, so a mostly-offloaded model is rarely worth using. Every figure here assumes it is fully resident on the card.

06

Would two GeForce RTX 4090 Max-Q cards be twice as fast?

Pairing GeForce RTX 4090 Max-Q cards buys headroom rather than pace: 32 GB of combined memory, at roughly the same generation speed as one.

07

What AI models can a GeForce RTX 4090 Max-Q run?

432 of the 679 open-weight language models we track fit on a GeForce RTX 4090 Max-Q and can be run locally on it. The table on this page lists every one, with the memory it needs, the quantisation it runs at and an estimated generation speed.

08

What is the largest AI model a GeForce RTX 4090 Max-Q can run?

The largest model in our catalogue that fits on a GeForce RTX 4090 Max-Q is Nemotron 3-Nano-30B-A3B at 31.6B parameters, compressed to Q3_K_M. It generates roughly 116 tokens per second and needs about 14.4 GB of the card's memory.

09

How many tokens per second does a GeForce RTX 4090 Max-Q produce?

It depends on the model. On a GeForce RTX 4090 Max-Q the fastest model we track is Gemma 3 QAT 1B at about 244 tokens per second, while larger models run proportionally slower because each token requires reading the whole model out of memory once. Speeds are estimates for a single conversation at a time.

10

Can a GeForce RTX 4090 Max-Q run a 7B model?

Yes. For example a GeForce RTX 4090 Max-Q runs Multi-Token Prediction 7B at Q8_0, using about 7.9 GB of memory and generating around 36.4 tokens per second.

11

Can a GeForce RTX 4090 Max-Q run a 13B model?

Yes. For example a GeForce RTX 4090 Max-Q runs DeepSeekMoE-16B at Q6_K, using about 13.7 GB of memory and generating around 123 tokens per second.

12

Can a GeForce RTX 4090 Max-Q run a 30B model?

Yes. For example a GeForce RTX 4090 Max-Q runs ERNIE-4.5-VL-28B-A3B at Q3_K_M, using about 12.9 GB of memory and generating around 131 tokens per second.

13

How much memory does a GeForce RTX 4090 Max-Q have?

A GeForce RTX 4090 Max-Q has 16 GB of GDDR6 memory. Around a tenth of that is reserved by the inference runtime and the driver, leaving roughly 14.4 GB available for a model and its conversation.

14

What is the memory bandwidth of a GeForce RTX 4090 Max-Q?

The GeForce RTX 4090 Max-Q has 576 GB/s of memory bandwidth, across a 256-bit memory bus. This is the single best predictor of how fast it generates text, because producing each token means reading the entire model out of memory once.

15

What type of memory does a GeForce RTX 4090 Max-Q use?

It uses GDDR6 clocked at 2.25 GHz. HBM types are found on datacentre accelerators and carry far more bandwidth than the GDDR used on desktop cards, which is why they generate tokens considerably faster at the same capacity.

16

Who makes the GeForce RTX 4090 Max-Q?

The GeForce RTX 4090 Max-Q is a NVIDIA product, with the chip manufactured by TSMC, on a 5 nm process.

17

When was the GeForce RTX 4090 Max-Q released?

The GeForce RTX 4090 Max-Q was released in January 2023.

18

How much power does a GeForce RTX 4090 Max-Q use?

The GeForce RTX 4090 Max-Q has a rated board power of 80 W. Generating text draws hard in bursts and idles between requests, so average consumption over a working session is normally well below the rated figure.

19

How much cache does a GeForce RTX 4090 Max-Q have?

The GeForce RTX 4090 Max-Q has 128 KB of L1 cache, and 64 MB of L2 cache. Cache absorbs part of the memory traffic that would otherwise reach the main bus, so a larger L2 gives a modest lift to generation speed beyond what bandwidth alone predicts.

20

What are the TFLOPS of a GeForce RTX 4090 Max-Q?

The GeForce RTX 4090 Max-Q is rated at 28.3 TFLOPS at half precision and 28.3 TFLOPS at single precision. These are peak arithmetic ceilings rather than achievable rates, and text generation reaches only a small fraction of them because it is limited by memory bandwidth instead.

The other direction

Looking at it from the other side?

This page starts from the hardware. If you already know which model you want and need to know what it takes to run it, start from the model instead.

All GPUs